Abstract
Presented at conference on science, technology, engineering and economics for Digital Agriculture (steeDA), University of Sydney, 3-5 Dec 2019
Physical and financial data from cooperating farmers near Corowa were analysed, including detailed geo-referenced header yield records from harvests between 2000 to 2017.
Historical long-term whole-farm average yields each year for two farms were used to calculate best-fit yield response curves against growing season rainfalls (GSR). R2 values for quadratic curves of wheat and canola were 76% and 84%, respectively, with both crops showing increasing yield depression at high GSR, caused by localised waterlogging.
Within-paddock variability with respect to GSR is illustrated in the cases of three paddocks for which yields of geo-referenced 90 x 90m grid areas were available over five or six years, including dry and wet seasons. Grid area responses to GSR differ, some yielding better than average in dry years but worse in wet, and others performing worse than average in the dry but best in the wet. Consequently, there was less than 5% chance that the first-year yield quartile for any individual grid area would be repeated over the next four years. Real-time NVDI appeared to be a more accurate indicator of future yield and profitability than either historical yield or EM38 measurements.
Replicated N-rate trials located in low and high EM38 zones on two farms showed no consistent yield response. Under these conditions the long-term, risk adjusted, break-even cost of implementing VRN would be met by less than 1% increase in yield, or 70% decrease in the amount of N fertiliser applied.
Physical and financial data from cooperating farmers near Corowa were analysed, including detailed geo-referenced header yield records from harvests between 2000 to 2017.
Historical long-term whole-farm average yields each year for two farms were used to calculate best-fit yield response curves against growing season rainfalls (GSR). R2 values for quadratic curves of wheat and canola were 76% and 84%, respectively, with both crops showing increasing yield depression at high GSR, caused by localised waterlogging.
Within-paddock variability with respect to GSR is illustrated in the cases of three paddocks for which yields of geo-referenced 90 x 90m grid areas were available over five or six years, including dry and wet seasons. Grid area responses to GSR differ, some yielding better than average in dry years but worse in wet, and others performing worse than average in the dry but best in the wet. Consequently, there was less than 5% chance that the first-year yield quartile for any individual grid area would be repeated over the next four years. Real-time NVDI appeared to be a more accurate indicator of future yield and profitability than either historical yield or EM38 measurements.
Replicated N-rate trials located in low and high EM38 zones on two farms showed no consistent yield response. Under these conditions the long-term, risk adjusted, break-even cost of implementing VRN would be met by less than 1% increase in yield, or 70% decrease in the amount of N fertiliser applied.
| Original language | English |
|---|---|
| Number of pages | 31 |
| Publication status | Published - 05 Dec 2019 |
| Event | steeDA: science, technology, engineering, and economics for Digital Agriculture Conference - The University of Sydney, Camperdown, Australia Duration: 03 Dec 2019 → 05 Dec 2019 https://sydney.edu.au/content/dam/corporate/documents/sydney-institute-of-agriculture/outreach-engagement/steeda-conference-draft.pdf |
Conference
| Conference | steeDA: science, technology, engineering, and economics for Digital Agriculture Conference |
|---|---|
| Abbreviated title | science technology engineering and economics for Digital Agriculture |
| Country/Territory | Australia |
| City | Camperdown |
| Period | 03/12/19 → 05/12/19 |
| Internet address |
